1
00:00:02,440 --> 00:00:06,200
Speaker 1: This is the last with your host JJ zacheris j

2
00:00:06,360 --> 00:00:14,400
Speaker 1: J Zacherison. What's up everyone, It's JJ zacharyesin in this

3
00:00:14,440 --> 00:00:18,239
Speaker 1: episode ten eighty six of the Late Round Fantasy Football Podcast.

4
00:00:18,960 --> 00:00:21,319
Speaker 1: Thanks for tuning in. All of you have heard me

5
00:00:21,320 --> 00:00:23,760
Speaker 1: talk about draft capital on this show, especially this time

6
00:00:23,800 --> 00:00:26,120
Speaker 1: of year when it comes to the ZAP model. When

7
00:00:26,120 --> 00:00:29,120
Speaker 1: it comes to forecasting prospects of the next level, draft

8
00:00:29,160 --> 00:00:32,400
Speaker 1: capital is the single most important input to consider, and

9
00:00:32,440 --> 00:00:34,920
Speaker 1: there's a lot of reasons for that. For one, better

10
00:00:34,960 --> 00:00:38,440
Speaker 1: players generally get drafted earlier. NFL teams they might be

11
00:00:38,479 --> 00:00:41,599
Speaker 1: stupid sometimes, but they're spending millions and millions of dollars

12
00:00:41,680 --> 00:00:44,040
Speaker 1: in an attempt to find the right prospect at each

13
00:00:44,159 --> 00:00:47,240
Speaker 1: draft pick. They're clearly going to be better at picking

14
00:00:47,240 --> 00:00:50,960
Speaker 1: those players than most. But also, when a team makes

15
00:00:50,960 --> 00:00:53,519
Speaker 1: an investment, that player is likely going to have more

16
00:00:53,560 --> 00:00:57,840
Speaker 1: immediate opportunity or opportunity in general. When we're playing fantasy football,

17
00:00:58,120 --> 00:01:01,360
Speaker 1: volume is everything, and if that player has higher draft capital,

18
00:01:01,520 --> 00:01:05,600
Speaker 1: there's a higher likelihood to see that volume. Now, clearly

19
00:01:05,640 --> 00:01:08,759
Speaker 1: it's not that simple, but those are two key reasons

20
00:01:08,840 --> 00:01:11,280
Speaker 1: as to why draft capital gets the signal that it does.

21
00:01:11,319 --> 00:01:14,920
Speaker 1: In models, and it's why it's important. But with all

22
00:01:14,920 --> 00:01:18,000
Speaker 1: that being said, with all this talk about draft capital

23
00:01:18,319 --> 00:01:20,640
Speaker 1: being the thing that we should care about most, you

24
00:01:20,800 --> 00:01:23,319
Speaker 1: probably think that it's close to the only thing to

25
00:01:23,360 --> 00:01:27,319
Speaker 1: care about. But that's not the case at all. On

26
00:01:27,400 --> 00:01:31,399
Speaker 1: today's episode, the Truth about Draft Capital, Why it's important,

27
00:01:31,680 --> 00:01:34,720
Speaker 1: but why it's not everything. Now, before I get to

28
00:01:34,720 --> 00:01:37,080
Speaker 1: the analysis, I wanted to send a quick reminder about

29
00:01:37,080 --> 00:01:40,400
Speaker 1: the Late Round Prospect Guide. It's launching in less than

30
00:01:40,440 --> 00:01:42,880
Speaker 1: a week. The guide is over one hundred and seventy

31
00:01:42,880 --> 00:01:45,720
Speaker 1: pages that covers everything that goes in the ZAP models.

32
00:01:45,880 --> 00:01:49,600
Speaker 1: There's rankings, there's prospect profile write ups for every running back,

33
00:01:49,640 --> 00:01:52,160
Speaker 1: wide receiver and tight end at the combine. There's year

34
00:01:52,240 --> 00:01:54,360
Speaker 1: two profile right ups for the guys entering their second

35
00:01:54,400 --> 00:01:57,200
Speaker 1: season in the NFL, and more. You can still pre

36
00:01:57,360 --> 00:01:59,800
Speaker 1: order it right now for fourteen ninety nine over on

37
00:01:59,880 --> 00:02:02,800
Speaker 1: lund dot com. In six days, that's gonna rise to

38
00:02:02,920 --> 00:02:07,560
Speaker 1: nineteen ninety nine, so get on that lateround dot com.

39
00:02:07,680 --> 00:02:10,400
Speaker 1: To help explain the truth about draft capital, I, of course,

40
00:02:10,800 --> 00:02:13,960
Speaker 1: when into the ZAP Moodel's database. That database has every

41
00:02:14,000 --> 00:02:16,560
Speaker 1: running back, wide receiver, and now tight end who is

42
00:02:16,600 --> 00:02:19,800
Speaker 1: either drafted or at the NFL combine since twenty eleven.

43
00:02:20,800 --> 00:02:23,359
Speaker 1: As I talked about on last week's Mailbag episode, the

44
00:02:23,480 --> 00:02:25,799
Speaker 1: ZAP model looks to predict the player's B two s,

45
00:02:26,080 --> 00:02:28,520
Speaker 1: or his best two seasons in PPR points per game

46
00:02:28,760 --> 00:02:31,760
Speaker 1: across his first three years in the league. Minimum seven

47
00:02:31,760 --> 00:02:33,520
Speaker 1: games played in a season for it to count, and

48
00:02:33,560 --> 00:02:35,760
Speaker 1: it throws out the final week of the regular season,

49
00:02:36,160 --> 00:02:39,760
Speaker 1: so we're only looking at fantasy relevant games to wrap

50
00:02:39,800 --> 00:02:41,880
Speaker 1: your head around B two s. Imagine a running back

51
00:02:41,960 --> 00:02:45,320
Speaker 1: scored ten, twelve, and fourteen PPR points per game over

52
00:02:45,360 --> 00:02:47,640
Speaker 1: his first three years, and he never missed a game,

53
00:02:48,280 --> 00:02:50,960
Speaker 1: His B two s would be thirteen. It would average

54
00:02:51,000 --> 00:02:54,520
Speaker 1: his best two seasons of twelve and fourteen. That's what

55
00:02:54,600 --> 00:02:57,760
Speaker 1: the ZAP model is attempting to forecast, is trying to

56
00:02:57,800 --> 00:03:01,400
Speaker 1: predict which players will score highest and be and it

57
00:03:01,480 --> 00:03:04,320
Speaker 1: gives a player some prospect score a ZAP score to

58
00:03:04,400 --> 00:03:07,240
Speaker 1: help with that. Now. I bring that up because when

59
00:03:07,240 --> 00:03:10,280
Speaker 1: talking about how important draft capital is, B two s

60
00:03:10,440 --> 00:03:13,040
Speaker 1: is how I'm gonna measure that importance. You can do

61
00:03:13,080 --> 00:03:15,960
Speaker 1: this in a ton of different ways. Best season PPR

62
00:03:16,040 --> 00:03:18,399
Speaker 1: points per game across his first three years, how many

63
00:03:18,480 --> 00:03:22,280
Speaker 1: top twelve top twenty four seasons the player had I

64
00:03:22,520 --> 00:03:25,880
Speaker 1: just do it via B two S, and since B

65
00:03:25,919 --> 00:03:28,919
Speaker 1: two S features three seasons of data on today's show,

66
00:03:29,160 --> 00:03:31,560
Speaker 1: I'm looking at players who are drafted from twenty eleven

67
00:03:31,840 --> 00:03:35,000
Speaker 1: through twenty twenty three. That way, they've had a chance

68
00:03:35,120 --> 00:03:38,240
Speaker 1: to play those three years. Now, I think it's human

69
00:03:38,320 --> 00:03:40,560
Speaker 1: nature for your brain to go into this linear mode

70
00:03:40,600 --> 00:03:43,720
Speaker 1: when we talk about draft capital. Maybe not totally linear,

71
00:03:43,760 --> 00:03:46,120
Speaker 1: but let me explain what I mean. You hear someone

72
00:03:46,120 --> 00:03:49,600
Speaker 1: say a wide receiver was drafted sixteenth overall. Then you

73
00:03:49,680 --> 00:03:52,080
Speaker 1: compare that player to someone who went thirty second overall,

74
00:03:52,240 --> 00:03:54,520
Speaker 1: and then you compare a third player who went forty

75
00:03:54,520 --> 00:03:58,119
Speaker 1: eighth overall. Those are all wide receivers who went sixteen

76
00:03:58,160 --> 00:04:02,720
Speaker 1: picks one after another. There were sixteen selections between each pick.

77
00:04:03,560 --> 00:04:07,480
Speaker 1: But performance versus draft capital doesn't behave in a linear way.

78
00:04:08,280 --> 00:04:10,360
Speaker 1: For running backs and wide receivers, the positions that we're

79
00:04:10,360 --> 00:04:13,200
Speaker 1: going to focus on today, it's a downward slope where

80
00:04:13,240 --> 00:04:15,600
Speaker 1: the first part of the first round that slope is

81
00:04:15,720 --> 00:04:18,800
Speaker 1: really steep, but then it starts to level out and

82
00:04:18,880 --> 00:04:23,440
Speaker 1: at some points draft capital is actually completely irrelevant. Now,

83
00:04:23,480 --> 00:04:25,760
Speaker 1: I want to start with running back first because quite frankly,

84
00:04:25,920 --> 00:04:28,200
Speaker 1: it's an easier position to model in general compared to

85
00:04:28,240 --> 00:04:31,520
Speaker 1: wide receiver. When it comes to big investments at running

86
00:04:31,520 --> 00:04:34,120
Speaker 1: back in the draft, they almost always give us good

87
00:04:34,120 --> 00:04:37,280
Speaker 1: fantasy production, which is why if Jeremiah loves draft capital

88
00:04:37,360 --> 00:04:40,120
Speaker 1: just keeps going up, he's actually gonna end up in

89
00:04:40,160 --> 00:04:43,159
Speaker 1: the legendary performer category in the model this year. I

90
00:04:43,160 --> 00:04:46,200
Speaker 1: didn't think that was gonna happen at first, but we

91
00:04:46,279 --> 00:04:48,400
Speaker 1: know that top half of the first round running backs

92
00:04:48,520 --> 00:04:51,160
Speaker 1: have been unreal over the last fifteen or so years

93
00:04:51,160 --> 00:04:54,440
Speaker 1: in fantasy football. Since twenty eleven, here are the guys

94
00:04:54,480 --> 00:04:58,680
Speaker 1: draft in the top sixteen. Trent Richardson, Melvin Gordon, Todd Gurley,

95
00:04:58,720 --> 00:05:02,599
Speaker 1: Ezekiel Elliot, Christian mcac affy, Leonard Fournette, Saquon Barkley, Bjeon Robinson,

96
00:05:02,680 --> 00:05:07,240
Speaker 1: Jamier Gibbs, and Ashton Genty and probably Jeremiah Love. All

97
00:05:07,320 --> 00:05:10,039
Speaker 1: of those guys scored roughly seventeen PPERO points per game

98
00:05:10,080 --> 00:05:12,240
Speaker 1: or more in one of their first three seasons, except

99
00:05:12,240 --> 00:05:15,800
Speaker 1: Gent who's only played one season. Even Trent Richardson got

100
00:05:15,839 --> 00:05:18,919
Speaker 1: there as a rookie. Needless to say, it's not surprising

101
00:05:19,040 --> 00:05:21,440
Speaker 1: that when you look at the correlation between draft capital

102
00:05:21,440 --> 00:05:23,800
Speaker 1: and B two S for first round running backs. There's

103
00:05:23,880 --> 00:05:28,360
Speaker 1: decent correlation. But let's take a step backwards first. Like

104
00:05:28,440 --> 00:05:31,320
Speaker 1: I said, this database has every running back who was

105
00:05:31,360 --> 00:05:35,119
Speaker 1: at the NFL Combine or drafted since twenty eleven. Anyone

106
00:05:35,160 --> 00:05:38,120
Speaker 1: who goes undrafted just gets a default draft capital of

107
00:05:38,240 --> 00:05:41,919
Speaker 1: undrafted their draft pick is all the same. The R

108
00:05:42,000 --> 00:05:45,160
Speaker 1: squared between draft capital, their raw pick, and B two

109
00:05:45,279 --> 00:05:48,560
Speaker 1: S is aboutzer point four to five, which means draft

110
00:05:48,600 --> 00:05:51,200
Speaker 1: capital explains roughly forty five percent of the variants in

111
00:05:51,279 --> 00:05:55,120
Speaker 1: B two S outcomes across this entire sample, which is

112
00:05:55,160 --> 00:05:59,480
Speaker 1: fairly strong correlation for fantasy football purposes. But that's across

113
00:05:59,520 --> 00:06:02,640
Speaker 1: the whole. I made a thread on Twitter last week

114
00:06:02,680 --> 00:06:05,000
Speaker 1: about prospecting through data, and one of the things that

115
00:06:05,040 --> 00:06:07,880
Speaker 1: I brought up in that thread was R squared analysis.

116
00:06:08,680 --> 00:06:11,640
Speaker 1: You hear data people like myself talk about R squared

117
00:06:11,680 --> 00:06:13,840
Speaker 1: all the time, and people will email me or hit

118
00:06:13,880 --> 00:06:16,080
Speaker 1: me up and they're like, hey, JJ, what's the R

119
00:06:16,200 --> 00:06:18,600
Speaker 1: squared of your model? Now, for those of you who

120
00:06:18,600 --> 00:06:21,039
Speaker 1: think I'm speaking a different language right now, are squared.

121
00:06:21,200 --> 00:06:23,479
Speaker 1: It's a statistical measure that tells us how much of

122
00:06:23,480 --> 00:06:26,880
Speaker 1: the variation in one variable can be explained by another variable.

123
00:06:27,360 --> 00:06:30,400
Speaker 1: When R squared is one, it means it can explain everything.

124
00:06:30,440 --> 00:06:33,960
Speaker 1: When it's zero, it means there's no linear relationship between

125
00:06:34,000 --> 00:06:37,240
Speaker 1: those two variables. So in this case, we're looking at

126
00:06:37,279 --> 00:06:40,320
Speaker 1: draft capital versus B two S, So we're seeing how

127
00:06:40,400 --> 00:06:43,040
Speaker 1: much of the variation in a player's NFL production can

128
00:06:43,080 --> 00:06:45,640
Speaker 1: be explained by a variable. In this case, that variable

129
00:06:45,720 --> 00:06:48,800
Speaker 1: is draft capital. But people will say, hey, JJ, what's

130
00:06:48,839 --> 00:06:51,479
Speaker 1: the R squared of your model? And my answer is

131
00:06:51,839 --> 00:06:54,800
Speaker 1: that doesn't matter without context. If I tell you that

132
00:06:54,839 --> 00:06:57,280
Speaker 1: the R squared from my model is point four two,

133
00:06:57,360 --> 00:06:59,239
Speaker 1: you might be like, oh, that's actually not that bad.

134
00:06:59,520 --> 00:07:02,719
Speaker 1: But actually that's really bad because that's worse than the

135
00:07:02,880 --> 00:07:06,760
Speaker 1: R squared of just plain old draft capital. And anyone

136
00:07:06,800 --> 00:07:10,360
Speaker 1: can improve that are squared by adding certain data points.

137
00:07:10,720 --> 00:07:14,160
Speaker 1: For example, if I removed all undrafted players, the R

138
00:07:14,200 --> 00:07:17,000
Speaker 1: square between draft capital and B two S lowers to

139
00:07:17,040 --> 00:07:20,000
Speaker 1: about point thirty nine. Now, just so we're all on

140
00:07:20,040 --> 00:07:22,520
Speaker 1: the same page, think about why that might be the

141
00:07:22,560 --> 00:07:26,280
Speaker 1: case is because I'm removing a bunch of undrafted players

142
00:07:26,400 --> 00:07:28,840
Speaker 1: who are almost definitely gonna do nothing in the NFL,

143
00:07:29,560 --> 00:07:32,560
Speaker 1: which means they have super late draft capital, they're undrafted,

144
00:07:32,800 --> 00:07:36,480
Speaker 1: and B two s's of zero one maybe two points.

145
00:07:37,440 --> 00:07:40,960
Speaker 1: That improves your correlation when they're in the data set.

146
00:07:41,880 --> 00:07:45,280
Speaker 1: Having the undrafted players in the model improves that correlation.

147
00:07:46,640 --> 00:07:50,960
Speaker 1: If you change that sample, then are squared changes. For example,

148
00:07:51,320 --> 00:07:54,400
Speaker 1: first round running backs. I told you that top half

149
00:07:54,400 --> 00:07:57,320
Speaker 1: first round running backs have been really good since twenty eleven,

150
00:07:57,360 --> 00:08:00,200
Speaker 1: and they have. That means that the back half after

151
00:08:00,200 --> 00:08:02,920
Speaker 1: the first round has crappier running backs. Then we'll see

152
00:08:02,960 --> 00:08:05,680
Speaker 1: some sort of correlation for first round running backs when

153
00:08:05,680 --> 00:08:08,960
Speaker 1: looking at draft capital versus B two S right, And

154
00:08:09,000 --> 00:08:12,520
Speaker 1: that's exactly what we see when looking only at first

155
00:08:12,600 --> 00:08:15,280
Speaker 1: round running backs in is twenty eleven, the R squared

156
00:08:15,440 --> 00:08:17,840
Speaker 1: between where they were drafted and how they performed in

157
00:08:17,880 --> 00:08:21,200
Speaker 1: B two S has been point five to six, which

158
00:08:21,240 --> 00:08:23,360
Speaker 1: is stronger than all of these running backs when you

159
00:08:23,400 --> 00:08:27,200
Speaker 1: look at round one through the undrafted guys. And that

160
00:08:27,240 --> 00:08:29,960
Speaker 1: should hint to us that later first round running backs

161
00:08:30,280 --> 00:08:32,880
Speaker 1: might not be that productive, or I should say they've

162
00:08:32,880 --> 00:08:35,520
Speaker 1: seen a dip in production compared to the guys drafted

163
00:08:35,520 --> 00:08:37,960
Speaker 1: in the front half of the first round, because if

164
00:08:38,000 --> 00:08:41,040
Speaker 1: they were just as productive, the correlation for first round

165
00:08:41,080 --> 00:08:44,280
Speaker 1: running backs wouldn't be very strong as you'd expect. As

166
00:08:44,320 --> 00:08:48,439
Speaker 1: you move through the draft round by round, that correlation weakens.

167
00:08:49,320 --> 00:08:52,120
Speaker 1: Running back is really top heavy. The top guys in

168
00:08:52,200 --> 00:08:55,680
Speaker 1: the position are doing so much work. They're almost always

169
00:08:55,720 --> 00:08:58,840
Speaker 1: hitting Round two running backs. In is twenty eleven, they've

170
00:08:58,840 --> 00:09:02,320
Speaker 1: had an R scored of points round three point one one,

171
00:09:02,760 --> 00:09:05,240
Speaker 1: And as you approach day three, you start to get

172
00:09:05,240 --> 00:09:08,360
Speaker 1: into numbers that are essentially just non existent where you

173
00:09:08,360 --> 00:09:10,640
Speaker 1: can just throw out draft capital and just pick who

174
00:09:10,679 --> 00:09:13,800
Speaker 1: you want. You can choose two end points here to

175
00:09:13,920 --> 00:09:17,160
Speaker 1: make your story, to create that narrative, like, for instance,

176
00:09:17,320 --> 00:09:19,560
Speaker 1: if you look at running backs, pick between picks seventeen

177
00:09:19,600 --> 00:09:22,560
Speaker 1: and fifty. Since twenty eleven through twenty twenty three, we've

178
00:09:22,559 --> 00:09:25,280
Speaker 1: had twenty five of them. You know what the correlation

179
00:09:25,640 --> 00:09:28,400
Speaker 1: between draft capital and B twos among those twenty five

180
00:09:28,440 --> 00:09:33,480
Speaker 1: running bax is again, pick seventeen to pick fifty, it's

181
00:09:33,640 --> 00:09:40,000
Speaker 1: point zero zero two. That's nonexistent correlation. What that tells

182
00:09:40,000 --> 00:09:42,520
Speaker 1: you is that historically there hasn't been much of a

183
00:09:42,520 --> 00:09:45,240
Speaker 1: difference in production for running backs drafted mid round one

184
00:09:45,559 --> 00:09:49,000
Speaker 1: through mid round two. But if you add in those

185
00:09:49,040 --> 00:09:51,720
Speaker 1: top sixteen running backs whoa all of a sudden, that

186
00:09:51,840 --> 00:09:56,560
Speaker 1: equation changes because remember those guys are doing a lot

187
00:09:56,559 --> 00:10:00,240
Speaker 1: of the heavy lifting. The R squared goes from non

188
00:10:00,320 --> 00:10:03,440
Speaker 1: existent to very existent. It shifts a point twenty five.

189
00:10:03,520 --> 00:10:07,480
Speaker 1: With thirty four running backs now in the sample, Like

190
00:10:07,520 --> 00:10:10,040
Speaker 1: I said, you can pick and choose your end points here. Like,

191
00:10:10,080 --> 00:10:13,360
Speaker 1: for instance, pick thirty three through pick one hundred, essentially

192
00:10:13,440 --> 00:10:17,560
Speaker 1: Day two picks the R squared point zero seven. That's

193
00:10:17,559 --> 00:10:20,800
Speaker 1: not very strong at all. Pick fifty to pick one hundred,

194
00:10:20,880 --> 00:10:24,720
Speaker 1: the R squared is zero point zero zero, non existent.

195
00:10:24,920 --> 00:10:28,080
Speaker 1: It doesn't matter. Now, this isn't to say that draft

196
00:10:28,080 --> 00:10:31,559
Speaker 1: capital is irrelevant for those players. All things being equal,

197
00:10:31,800 --> 00:10:33,760
Speaker 1: you'd still probably want the player who goes earlier in

198
00:10:33,800 --> 00:10:37,000
Speaker 1: the draft. And again, I'm just choosing end points here.

199
00:10:37,440 --> 00:10:40,360
Speaker 1: You can capture different players at different draft capitals and

200
00:10:40,440 --> 00:10:43,520
Speaker 1: make it say one thing or another, but one thing

201
00:10:43,520 --> 00:10:46,440
Speaker 1: here is super clear. The further you move through the draft,

202
00:10:46,679 --> 00:10:50,520
Speaker 1: the less draft capital matters. Now, another way to approach

203
00:10:50,559 --> 00:10:53,760
Speaker 1: this topic is through binning. We can just take all

204
00:10:53,760 --> 00:10:55,920
Speaker 1: the Round two running backs and group them together, and

205
00:10:55,920 --> 00:10:57,480
Speaker 1: then we can do the exact same thing with Round

206
00:10:57,520 --> 00:11:00,000
Speaker 1: three running backs, and then we can just compare those

207
00:11:00,080 --> 00:11:03,520
Speaker 1: two groups, and I think that's probably the analysis that

208
00:11:03,559 --> 00:11:06,439
Speaker 1: you see most And there's a chance that it's contradicting

209
00:11:06,679 --> 00:11:09,840
Speaker 1: what you're currently hearing me talk about, but it's not

210
00:11:10,000 --> 00:11:13,120
Speaker 1: because binning does a different job than our squared analysis does.

211
00:11:13,600 --> 00:11:17,160
Speaker 1: Our square to saying how smooth this production shift from

212
00:11:17,200 --> 00:11:20,480
Speaker 1: one draft pick to another. When we're binning, the group

213
00:11:20,520 --> 00:11:23,480
Speaker 1: of players within that bin can still be pretty chaotic.

214
00:11:24,440 --> 00:11:26,439
Speaker 1: Like if we were to compare round two to round

215
00:11:26,440 --> 00:11:29,040
Speaker 1: three players. That means that we're looking at all players

216
00:11:29,120 --> 00:11:31,760
Speaker 1: drafted between pick thirty three and sixty four and we're

217
00:11:31,800 --> 00:11:35,040
Speaker 1: grouping them together. But the players within that group could

218
00:11:35,040 --> 00:11:38,680
Speaker 1: produce in wildly different ways, like a player pick sixty

219
00:11:38,679 --> 00:11:42,160
Speaker 1: fourth can outperform a player pick thirty third, But when

220
00:11:42,160 --> 00:11:46,280
Speaker 1: they're bin together, that's not discovered when we group them up,

221
00:11:46,880 --> 00:11:50,240
Speaker 1: because they're all just one group. When you approach this

222
00:11:50,280 --> 00:11:53,599
Speaker 1: type of analysis via binning, you'll likely find some differences.

223
00:11:54,280 --> 00:11:57,240
Speaker 1: For instance, the average B twos among round two running

224
00:11:57,280 --> 00:12:00,360
Speaker 1: backs has been about ten point seven. For round three,

225
00:12:00,520 --> 00:12:04,439
Speaker 1: it falls to nine point six. That's not amazingly significant,

226
00:12:04,640 --> 00:12:08,280
Speaker 1: but it's significant enough. Meanwhile, running backs get fourteen or

227
00:12:08,320 --> 00:12:10,600
Speaker 1: more PPR points in B two s at a thirty

228
00:12:10,600 --> 00:12:13,800
Speaker 1: five percent rate out of round two. Out of round three,

229
00:12:14,080 --> 00:12:16,719
Speaker 1: that's twenty two percent. They get to twelve or more

230
00:12:16,760 --> 00:12:19,240
Speaker 1: points at a forty four percent rate versus a twenty

231
00:12:19,280 --> 00:12:22,880
Speaker 1: seven percent rate. You get the picture. In general, a

232
00:12:22,920 --> 00:12:25,760
Speaker 1: player drafted in round two is gonna be better than

233
00:12:25,800 --> 00:12:28,720
Speaker 1: a player drafted in round three. But those are also

234
00:12:28,840 --> 00:12:32,920
Speaker 1: pretty arbitrary thresholds, and far too often, way too often,

235
00:12:33,160 --> 00:12:35,679
Speaker 1: you get people looking at players who are fifteen to

236
00:12:35,720 --> 00:12:39,000
Speaker 1: twenty picks apart, and they automatically put that better draft

237
00:12:39,000 --> 00:12:42,880
Speaker 1: capital player ahead in their rankings. And it's because of

238
00:12:42,960 --> 00:12:46,800
Speaker 1: draft capital, but realistically it's not that significant in the

239
00:12:46,800 --> 00:12:50,079
Speaker 1: grand scheme of things. Really, the takeaway here is that

240
00:12:50,200 --> 00:12:52,600
Speaker 1: draft capital works more in tiers than it does as

241
00:12:52,640 --> 00:12:56,120
Speaker 1: a precise ranking system, because when you treat draft capital

242
00:12:56,160 --> 00:12:59,400
Speaker 1: too literally, you end up viewing pick sixty four as

243
00:12:59,520 --> 00:13:03,160
Speaker 1: meaningful better than pick sixty five. And when you look

244
00:13:03,200 --> 00:13:05,720
Speaker 1: at the RS squared among Day two running backs, it

245
00:13:05,840 --> 00:13:08,280
Speaker 1: tells us that production in that range is far more

246
00:13:08,360 --> 00:13:10,760
Speaker 1: random than that. And then as you get to day

247
00:13:10,760 --> 00:13:14,079
Speaker 1: three there's some correlation that does exist. Picks one hundred

248
00:13:14,120 --> 00:13:16,040
Speaker 1: plus to the end of the draft, it's about point

249
00:13:16,120 --> 00:13:19,520
Speaker 1: one five, which does mean a fourth rounder, for instance,

250
00:13:19,679 --> 00:13:21,760
Speaker 1: has a little bit more likelihood to hit than a

251
00:13:21,800 --> 00:13:25,319
Speaker 1: seventh rounder, But when you get into round five, round six,

252
00:13:25,880 --> 00:13:28,760
Speaker 1: there's really no difference between those types of players and

253
00:13:28,800 --> 00:13:32,240
Speaker 1: a round seven running back. The TLDR here is at

254
00:13:32,240 --> 00:13:34,920
Speaker 1: the top of running back draft classes, Oftener doing the

255
00:13:34,960 --> 00:13:37,640
Speaker 1: heavy lifting. When it comes to draft capital being a

256
00:13:37,720 --> 00:13:41,280
Speaker 1: huge point of emphasis, you should almost blindly follow and

257
00:13:41,360 --> 00:13:43,600
Speaker 1: just draft top half of the first round running backs

258
00:13:43,880 --> 00:13:46,200
Speaker 1: because they hit at such a ridiculous rate in fantasy.

259
00:13:46,800 --> 00:13:51,000
Speaker 1: After that, don't be too certain about players. Make sure

260
00:13:51,000 --> 00:13:54,600
Speaker 1: you're looking deeper than just draft capital. Even at running back,

261
00:13:54,720 --> 00:13:56,960
Speaker 1: where I think the consensus believes that draft capital is

262
00:13:57,040 --> 00:14:01,120
Speaker 1: absolutely tied to production no matter what, that's just not

263
00:14:01,240 --> 00:14:04,560
Speaker 1: the case. Now. The same general trends can be found

264
00:14:04,559 --> 00:14:07,240
Speaker 1: at wide receiver, except it's not as clear cut. It's

265
00:14:07,280 --> 00:14:10,640
Speaker 1: a little bit more wild wild west. Across the entire

266
00:14:10,640 --> 00:14:13,120
Speaker 1: sample of players, we see an R squared of point

267
00:14:13,160 --> 00:14:15,600
Speaker 1: four to eight. Technically, that's better than what we see

268
00:14:15,600 --> 00:14:18,400
Speaker 1: at running back. Again, that's looking at all round one

269
00:14:18,840 --> 00:14:22,120
Speaker 1: through undrafted players since twenty eleven, but there's a lot

270
00:14:22,160 --> 00:14:25,320
Speaker 1: more nuance to the wide receiver stuff. Remember how first

271
00:14:25,360 --> 00:14:28,320
Speaker 1: round running backs saw really strong correlation because the top

272
00:14:28,320 --> 00:14:30,960
Speaker 1: of the position has just been so good. Well, since

273
00:14:31,000 --> 00:14:33,600
Speaker 1: twenty eleven, that hasn't been the case. At wide receiver.

274
00:14:34,200 --> 00:14:37,040
Speaker 1: The R squared for first round wide outs is just

275
00:14:37,160 --> 00:14:40,560
Speaker 1: point zero eight. Late first wide outs have been decent

276
00:14:40,560 --> 00:14:44,360
Speaker 1: compared to early first round ones. Now. Fortunately, it seems

277
00:14:44,360 --> 00:14:46,200
Speaker 1: like teams have gotten a little bit better about this

278
00:14:46,240 --> 00:14:49,400
Speaker 1: in recent years, but it's still not nearly as strong

279
00:14:49,480 --> 00:14:52,000
Speaker 1: as it is at running back. When running backs get

280
00:14:52,040 --> 00:14:57,359
Speaker 1: drafted early, they almost always produce. There's literally zero correlation

281
00:14:57,560 --> 00:15:00,360
Speaker 1: for round two wide receivers, and then it's the same

282
00:15:00,400 --> 00:15:03,400
Speaker 1: for round three, And when you combine the two, the

283
00:15:03,520 --> 00:15:06,880
Speaker 1: R squared is point zero five, which at face value

284
00:15:07,200 --> 00:15:09,640
Speaker 1: tells you that there's very little correlation when it comes

285
00:15:09,680 --> 00:15:13,200
Speaker 1: to day two wide receivers in general. With that being said,

286
00:15:13,560 --> 00:15:17,160
Speaker 1: round two wide receivers do outperform Round three wide receivers.

287
00:15:17,640 --> 00:15:20,320
Speaker 1: Their B two s is about two point three points higher.

288
00:15:20,960 --> 00:15:23,520
Speaker 1: They hit ten, twelve and fourteen p Piero points in

289
00:15:23,560 --> 00:15:27,000
Speaker 1: B two s at higher rates across the board, but

290
00:15:27,160 --> 00:15:31,560
Speaker 1: those rates aren't unbelievably significant, and the average for the

291
00:15:31,640 --> 00:15:34,360
Speaker 1: round two guys, it usually comes down to floor rather

292
00:15:34,400 --> 00:15:37,520
Speaker 1: than ceiling, and in fantasy football, ceiling is typically the

293
00:15:37,560 --> 00:15:40,480
Speaker 1: thing that we're chasing. But this continues to play into

294
00:15:40,480 --> 00:15:44,360
Speaker 1: this idea that draft capital isn't linear and round selection

295
00:15:44,480 --> 00:15:47,720
Speaker 1: is still relatively arbitrary. Just because a player goes thirty

296
00:15:47,720 --> 00:15:49,480
Speaker 1: third overall doesn't mean that you should take him in

297
00:15:49,520 --> 00:15:52,040
Speaker 1: your rookie draft, no matter what, over someone who goes

298
00:15:52,080 --> 00:15:55,360
Speaker 1: fifty third overall. In fact, this data is telling you

299
00:15:55,560 --> 00:15:57,880
Speaker 1: that there historically hasn't been that much of a difference

300
00:15:57,880 --> 00:16:01,440
Speaker 1: between those two players at wide receiver. Now, the interesting

301
00:16:01,480 --> 00:16:03,520
Speaker 1: thing is when you look at draft capital from twenty

302
00:16:03,520 --> 00:16:05,760
Speaker 1: eleven to twenty twenty three and you find the R

303
00:16:05,800 --> 00:16:08,080
Speaker 1: squared for players drafted in the top one hundred versus

304
00:16:08,160 --> 00:16:10,840
Speaker 1: B two s and then players drafted on Day three,

305
00:16:11,120 --> 00:16:14,000
Speaker 1: which I'm just signaling as anyone drafted after pick one

306
00:16:14,040 --> 00:16:16,880
Speaker 1: hundred because it's just cleaner. Keep in mind, for all

307
00:16:16,960 --> 00:16:20,160
Speaker 1: drafted players that are squared is point three to nine,

308
00:16:21,280 --> 00:16:24,880
Speaker 1: very similar to running back. We're removing the undrafted guys,

309
00:16:25,520 --> 00:16:28,280
Speaker 1: but for top one hundred picks at running back, draft

310
00:16:28,320 --> 00:16:31,160
Speaker 1: slot explains roughly fifty percent more of the B two

311
00:16:31,320 --> 00:16:35,040
Speaker 1: s variants than it does at wide receiver, meaning running

312
00:16:35,040 --> 00:16:38,560
Speaker 1: back correlation with top one hundred picks is much much stronger,

313
00:16:39,320 --> 00:16:42,040
Speaker 1: and that shouldn't be that surprising because of what running

314
00:16:42,040 --> 00:16:44,160
Speaker 1: back is doing in the top half of the first round.

315
00:16:45,080 --> 00:16:47,080
Speaker 1: Not to be fair, things have gotten a little bit

316
00:16:47,120 --> 00:16:49,800
Speaker 1: better at wide receiver during the twenty eighteen to twenty

317
00:16:49,840 --> 00:16:52,440
Speaker 1: twenty three time frame. Within that range, things have gotten

318
00:16:52,480 --> 00:16:55,400
Speaker 1: a little bit better. But the reason the wide receiver

319
00:16:55,480 --> 00:16:58,840
Speaker 1: correlation across the entire sample is just as strong as

320
00:16:58,880 --> 00:17:01,200
Speaker 1: it is at running back but it's so much weaker

321
00:17:01,400 --> 00:17:04,800
Speaker 1: with top one hundred picks is because wide receiver is

322
00:17:04,920 --> 00:17:08,080
Speaker 1: not very good on Day three. Day three running backs

323
00:17:08,080 --> 00:17:09,880
Speaker 1: at least will give us something a little bit more

324
00:17:09,920 --> 00:17:14,000
Speaker 1: frequently that can add more variance to the R squared analysis.

325
00:17:14,920 --> 00:17:17,480
Speaker 1: But you'll often get a lot of zeros from Day

326
00:17:17,480 --> 00:17:21,920
Speaker 1: three wide receivers, and when guys hit, they'll typically really hit,

327
00:17:22,080 --> 00:17:24,640
Speaker 1: but there's not enough of them to really drive down

328
00:17:24,640 --> 00:17:28,200
Speaker 1: that R squared. But look, we've talked a lot about

329
00:17:28,200 --> 00:17:31,000
Speaker 1: these different areas of the draft and are squared and binning.

330
00:17:31,280 --> 00:17:33,960
Speaker 1: What am I really getting at here? Well? I think

331
00:17:34,000 --> 00:17:36,399
Speaker 1: people need to know that while draft capital is really

332
00:17:36,440 --> 00:17:39,639
Speaker 1: really important. If you're choosing between two players who are

333
00:17:39,680 --> 00:17:42,760
Speaker 1: in the vicinity of one another, don't let draft capital

334
00:17:43,160 --> 00:17:47,160
Speaker 1: overtake your decision making. The NFL is better at identifying

335
00:17:47,240 --> 00:17:51,680
Speaker 1: tiers than ranking players within those tiers. Teams can identify

336
00:17:51,720 --> 00:17:55,600
Speaker 1: which players belong in some draft range, but the exact

337
00:17:55,800 --> 00:17:59,840
Speaker 1: order of those players is actually really really noisy. So

338
00:18:00,040 --> 00:18:01,960
Speaker 1: if you like a player who has picked fifty fourth

339
00:18:02,040 --> 00:18:05,120
Speaker 1: overall more than a player who has picked thirty fourth overall,

340
00:18:05,960 --> 00:18:08,240
Speaker 1: go with the guy who has picked fifty fourth overall

341
00:18:08,240 --> 00:18:11,080
Speaker 1: on your rookie draft. Not only is it more fun

342
00:18:11,320 --> 00:18:14,600
Speaker 1: to follow your own evaluations, but matth actually backs you

343
00:18:14,720 --> 00:18:19,119
Speaker 1: up with that move. Realistically, small draft capital differences are

344
00:18:19,240 --> 00:18:23,159
Speaker 1: usually meaningless. A fifteen to twenty pick difference inside the

345
00:18:23,160 --> 00:18:28,119
Speaker 1: same tier historically hasn't explained much about player outcomes, and

346
00:18:28,200 --> 00:18:33,119
Speaker 1: keep in mind round boundaries they're pretty arbitrary. Pick thirty

347
00:18:33,119 --> 00:18:36,280
Speaker 1: three versus pick thirty five, versus pick forty versus pick fifty.

348
00:18:36,960 --> 00:18:39,600
Speaker 1: They all often behave more similarly than the round two

349
00:18:39,680 --> 00:18:43,200
Speaker 1: or round three labels imply, which is precisely why I've

350
00:18:43,200 --> 00:18:47,480
Speaker 1: always used raw draft pick within the ZAP model. Now, look,

351
00:18:47,480 --> 00:18:49,480
Speaker 1: if you want this position. By position, I'd say the

352
00:18:49,520 --> 00:18:52,400
Speaker 1: following top half of the first round running backs. They're

353
00:18:52,440 --> 00:18:55,840
Speaker 1: extremely safe bets and they drive a large portion of

354
00:18:55,840 --> 00:18:59,800
Speaker 1: the draft capitol's predictiveness power at the position. Running back

355
00:19:00,280 --> 00:19:05,240
Speaker 1: starts to flatten after that top sixteen, top twenty range. Meanwhile,

356
00:19:05,240 --> 00:19:07,680
Speaker 1: at wide receiver, the evaluation with draft capital is a

357
00:19:07,680 --> 00:19:10,840
Speaker 1: little bit looser. Day two wide receivers often behave like

358
00:19:10,880 --> 00:19:14,159
Speaker 1: a large talent pool rather than some ranked list. You

359
00:19:14,160 --> 00:19:16,560
Speaker 1: should generally prefer a player drafted in round two versus

360
00:19:16,640 --> 00:19:19,199
Speaker 1: round three, but again, you don't need to force the

361
00:19:19,240 --> 00:19:24,440
Speaker 1: issue with arbitrary thresholds. Great wide receivers they're found throughout

362
00:19:24,480 --> 00:19:28,360
Speaker 1: Day two lead with your analysis. You shouldn't just rank

363
00:19:28,440 --> 00:19:32,080
Speaker 1: these guys based on draft capital because it's not nearly

364
00:19:32,119 --> 00:19:35,399
Speaker 1: as linear as you think. And that's just the truth.

365
00:19:36,560 --> 00:19:38,119
Speaker 1: That's it for today's show, though, thanks to all of

366
00:19:38,119 --> 00:19:40,000
Speaker 1: you for listening. If you had subscribed to the Late

367
00:19:40,040 --> 00:19:42,919
Speaker 1: Round Fantasy Football podcast, make sure you are by stitching

368
00:19:42,960 --> 00:19:45,399
Speaker 1: for it pretty much anywhere podcast can be found, and

369
00:19:45,760 --> 00:19:48,080
Speaker 1: follow me on Twitter and on Blue Sky at Late

370
00:19:48,160 --> 00:19:51,920
Speaker 1: Round QB. I'll be back on Wednesday live on YouTube

371
00:19:52,119 --> 00:19:54,760
Speaker 1: with a new Late Round Fantasy Football show and it's

372
00:19:54,840 --> 00:19:57,680
Speaker 1: just me. I'm doing a solo show because David Kitchen

373
00:19:57,720 --> 00:20:00,600
Speaker 1: is gone this week on vacation. After that, later in

374
00:20:00,640 --> 00:20:03,040
Speaker 1: the week, I've got a late Round Perspective episode for

375
00:20:03,040 --> 00:20:06,600
Speaker 1: you guys. Thanks for listening, greatly appreciate all of you.

376
00:20:07,000 --> 00:20:07,840
Speaker 1: I'll talk to you soon.